Download icon
Index
Download icon
Tweet
Download icon
Download Chapter
Download icon
Method
Download Single
Chapter

5

Research

Non-participant approach

Non-participant observation

Researchers collect data by observing behavior without actively interacting with the participants.

01 You can also do overt non-participant observation, for example, when researchers sit in on meetings or workshops on site, but do not actively participate. See also the textbox in #TISDD called Overt vs. covert research in 5.1.3.

In contrast to participant observation, researchers take a more distant role in non-participant approaches and do not interact with the research subjects; they behave like a “fly on the wall.” [01] Research subjects are often customers, employees, or other stakeholders, observed in situations that are relevant to the research question, such as using or providing a service or product, whether physical or digital. Often, non-participant observation is used to level out researcher biases in other methods and to reveal differences between what people say and what they actually do.

Non-participant observation can be overt or covert. Overt means that research subjects know that researchers are present, but they do not interact with each other – for example, when a researcher joins employees for meetings without interfering at all. This can be combined with other methods, like in-depth interviews to debrief afterwards and learn the different perspectives and hidden agendas of people attending the meeting. Overt non-participant observation can be biased through the observer effect, when people change or seek to improve an aspect of their behavior just because they are aware of being observed. Covert non-participant observation refers to observing research subjects without them knowing that they are being observed at all. Sometimes researchers pretend to be customers or passers-by, or even use one-way mirrors, for example. Covert non-participant observation minimizes the risk of people being affected by the presence of a researcher. Setting aside potential ethical concerns, it is also often the method of choice if people are unwilling to participate in your research.

During non-participant observations, it is important to observe not only what people are doing (for example, by interpreting their body language and gestures), but also what people are not doing (perhaps ignoring instructions or refraining from asking for help or assistance). Depending on the country and organization you’re working with, do not forget to check what kind of legal, ethical, and confidentiality agreements you need in advance and which forms of data you are allowed to collect, particularly in covert non-participant observations. Avoid taking photos or videos of strangers without their consent. If you cannot take photos or videos, use sketching or reconstruct the situation with a colleague afterwards to capture the situational context.

Duration
Preparation: 0.5 hours–2 weeks (depending on accessibility and legal regulations) // Activity: 1 hours–4 weeks (depending on number of observations and research objective) // Follow-up: 0.5 hours–2 weeks (depending on amount of data and collected data types)
Physical requirements
Notebook, photo camera, video camera, voice recorder, legal agreements (consent and/or confidentiality agreement)
Energy level
Middle
Researchers/Facilitators
Minimum 1 (it’s better to have 2–3 researchers)
Participants
Minimum 5 (but aim for at least 20 per group)
Expected output
Text (field notes), photos, videos, audio recordings, sketches, artifacts, statistics (e.g., counting customers per hour)
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Try to differentiate between concrete ­observations and your own interpretations ­(first-level/­second-level constructs).
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Try to differentiate between concrete ­observations and your own interpretations ­(first-level/­second-level constructs).
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Often there’s a difference between what ­people say and what people do. Use triangulation to ­cross-check your findings between methods.
Try to differentiate between concrete ­observations and your own interpretations ­(first-level/­second-level constructs).

Step-by-step guide

  1. Specify research question Define your research question or the focus of what you are interested in. Consider why you are doing research (exploratory vs. confirmatory research), what you want to do with your findings (personas, journey maps, system maps, etc.), and what sample size you’ll probably need. 
  2. Plan and prepare Based on your research question, define criteria for selecting suitable locations and situations for your non-participant observation. Depending on the research focus it might be more important to think about whom you observe and in what situation, or it might be more important to focus on the situational context: the when and where. Think about what types of data you are allowed to collect and if you’ll do overt or covert non-participant observation. Also, consider who you want to include as researchers from the client side or from other departments involved in the project. Summarize this in some brief observation guidelines based on what you want to find out, how you will do this, and what you aim to do with the data. 
  3. Conduct observations During non-participant observation, try to interfere with the research subjects as little as possible. Using a smartphone or any other unobtrusive device to collect your data might help. You can mix non-participant observation with other methods such as in-depth (retrospective) interviews afterwards to debrief observed situations. During your observations, try to collect as much unbiased “first-level construct” raw data as possible. The length and depth of participant observations varies with the research objective: from many quick, two-minute observations at a specific moment in a customer journey to observations of several days or sometimes even weeks – for example, when you do overt non-participant observations of a project team over the entire project duration.
  4. Follow-up Write up your individual key learnings from the observations right afterwards and compare them within your team. Review all your data and index it; highlight important passages. Try to find patterns within your data. For each non-participant observation session, write a short summary that includes your conflated key findings as well as raw data to exemplify these, such as quotes, photos, or videos. Don’t forget to link the summary to your underlying data by using indices.

Method notes

  • ‍Besides obvious qualitative research, such as observing body language, gestures, flow, usage of space or artifacts, interactions, and the like, researchers can also do some quantitative research, such as counting (a) how many customers within the hour pass by a shop, (b) how many of these come into the shop, and (c) how many of these start interacting with employees. The numbers can be aggregated to a simple conversion funnel, (a) → (b) → (c), and compared with data from other shops or other channels, like an online conversion funnel. In this context, researchers can observe situations with other people, digital interfaces, or machines.
  • A rather special approach to carrying out non-participant observation is call monitoring: researchers listening to phone calls. This is mostly used in call centers to research conversations between call-center agents and customers. Call monitoring can be done live or based on recorded phone calls. Conversations can then be analyzed to understand common problems of both customers and employees. Today, augmented-reality headsets, wearable sensors, and other recording devices are providing service designers with new data-collection methods that present fresh avenues of inquiry, as well as emerging demands for privacy and consent-management discipline. 
End of
Method
Non-participant observation
Taken from #TiSDD
Chapter
5
Research
Our BACKGROUND